随机风场覆冰导线舞动单模态模型的分段稀疏辨识对比研究

A Comparative Study on Piecewise Sparse Identification of Iced Bundle Conductor Galloping in Random Wind Field

覆冰输电线路舞动严重威胁着电力系统运行的安全性和稳定性.但是由于覆冰形状和风的随机性,对于实际覆冰导线舞动的数学模型建立目前没有实用的方法.本文基于数据驱动稀疏辨识算法,提出随机风荷载作用下覆冰四分裂导线舞动模型的辨识方法.首先基于Hamilton原理推导覆冰四分裂导线的动力学偏微分方程,再采用Galerkin法得到覆冰四分裂导线的动力学微分方程,并引入由Davenport谱生成、子段线性插值处理的随机风气动模型,继而获得随风速时变的覆冰四分裂导线舞动方程模型.最后结合不同的数据处理方法提出子段线性插值积分辨识法和子段线性插值微分辨识法,并应用到覆冰四分裂导线舞动均值模型辨识.通过一百组平均风速为10~30 m/s的计算机实验,探究并比较两种方法的辨识精度、辨识效率以及辨识稳定性.结果发现:随平均风速的变化,除位移三次项外,两种方法对覆冰四分裂导线舞动均值模型均有良好的辨识精度,且从响应的相对误差、辨识精度与稳定性以及辨识效率分析,微分辨识法更加优于积分辨识法,尤其对于速度一次项和三次项系数.本文研究成果可为输电线舞动模型的建立提供一定参考.

The galloping of iced transmission lines seriously threatens the safety and stability of the operation of the power system.However,due to the randomness of the ice shape and the wind,there is currently no practical way to establish a mathematical model of actual iced conductor galloping.Based on the data-driven sparse recognition algorithm,this paper proposes an identification method for the galloping model of iced quad bundle conductor under random wind loading.Firstly,the dynamic partial differential equation of the iced quad bundle conductor is derived based on Hamilton's principle,and then the Galerkin method is used to obtain the dynamic differential equation of the iced quad bundle conductor.The random wind aerodynamic model generated by the Davenport spectrum and processed by the linear interpolation of the sub-segments is introduced,and then the galloping equation model of iced quad bundle conductor with time-varying wind speed is obtained.Finally,combined with different data processing methods,the sub-segment linear interpolation integral recognition method and the sub-segment linear interpolation differential recognition method are proposed,and applied to the iced quad bundle conductor galloping mean model recognition.Through 100 sets of computer experiments with an average wind speed of 10~30m/s,the recognition accuracy,recognition efficiency and recognition stability of the two methods were explored and compared.The results show that with the change of average wind speed,except for the displacement cubic term,the two methods have good recognition accuracy for the iced quad bundle conductor galloping mean model.From the viewpoints of relative error of response,recognition accuracy and stability,and recognition efficiency,the differential recognition method is better than the integral recognition method,especially for the primary term and the third term coefficients of velocity.The results in this paper can provide reference for the establishment of transmission line galloping model.